SciQAG: A Framework for Auto-Generated Science Question Answering Dataset with Fine-grained Evaluation
This addresses the need for large-scale, high-quality datasets to evaluate and enhance science question-answering abilities in AI systems, particularly for researchers in natural language processing and scientific domains.
The authors tackled the problem of generating high-quality science question-answer pairs from scientific literature by introducing SciQAG, a framework using large language models, resulting in a dataset of 188,042 QA pairs from 22,743 papers across 24 domains and showing that fine-tuning LLMs on it significantly improves performance on open-ended QA and scientific tasks.
We introduce SciQAG, a novel framework for automatically generating high-quality science question-answer pairs from a large corpus of scientific literature based on large language models (LLMs). SciQAG consists of a QA generator and a QA evaluator, which work together to extract diverse and research-level questions and answers from scientific papers. Utilizing this framework, we construct a large-scale, high-quality, open-ended science QA dataset containing 188,042 QA pairs extracted from 22,743 scientific papers across 24 scientific domains. We also introduce SciQAG-24D, a new benchmark task designed to evaluate the science question-answering ability of LLMs. Extensive experiments demonstrate that fine-tuning LLMs on the SciQAG dataset significantly improves their performance on both open-ended question answering and scientific tasks. To foster research and collaboration, we make the datasets, models, and evaluation codes publicly available, contributing to the advancement of science question answering and developing more interpretable and reasoning-capable AI systems.